How to Beat a Chess Computer – Part 1
In order to claim victory against a chess computer, you’ll need to understand which aspects of playing games computers are superior at over humans, an ability to predict what they might not be able to work out, and a general understanding of how they think.
Since Gary Kasparov’s defeat at the hands of Deep Blue, it has become clear that computers have a knack for playing chess. Fortunately, they have flaws, which provide you with an opportunity to improve your chances of winning if you put in some preparation time. A chess computer works out who is winning differently than many humans would. There is no difference to the value of the pieces but only some humans would take into account a number of the factors that computers do.
How a chess computer plays
There’s a simple method for playing perfect chess: note down all the possible games and whether the final position is either won, lost, or drawn. Then work backwards while assuming that each player chooses the best line. After you’ve finished this process, you’ll know each of the best possible games. This isn’t a workable strategy for even today’s strongest computers, however. Therefore, a compromise is in order.
Instead, you would expand the list as large as possible, within the given time, and then apply an evaluation function in a bid to determine the probable outcome from each variation’s final position. Computers make use of evaluation functions in order to predict the probable result of the game from that position. It doesn’t, however, need to use much tactical information, as the tactics in the position are worked out within the actual variations.
Therefore, the computer’s strategic knowledge of the game is contained with the evaluation function. Commercial chess programs generally keep their evaluation function secret but, generally, it takes into account a number of factors such as Material, Development, and centralisation. The computer takes your score and subtracts it from its own. If the resulting score is positive, the computer is winning. If the score is negative, the human is winning.
How a chess computer works out variations
The compeer may well be putting in a lot of work when it comes to evaluating single positions. In order for the evaluation function to play practical chess, however, it requires more sophistication. Therefore, it doesn’t require as much time to calculate the possible variations, and there may well be less tactical ability due to a more advanced evaluation function.
There are numerous variations in chess positions, and the greater number of possible positions we find, the more time is needed in assessing them. It’s important to recognise quickly when a certain isn’t worth any further time being spent on it.
Mainly, the computers apply techniques based on ordering moves, so the one most likely to succeed is attempted first. Having determined lines without much potential, it will spend less time on assessing these. Therefore, it may not look at all possible sacrifices in much depth.